Passing Control Args

Steven M. Mortimer

2020-07-08

If you’re inserting records from R you may want to turn off the assignment rules or even bypass duplicate rules and alerts to save records. Beginning in Version 0.1.3 of the {salesforcer} package many functions have a control argument that will allow you to fine tune the behavior of calls to the Salesforce APIs. This vignette will introduce the different options you can control and how to pass them into the {salesforcer} functions you’re already familiar with.

The new control argument

This new feature can be seen in the sf_create (and many other functions) as control=list(...). The dots mean that you can pass any number of controls directly into the function.

First, authenticate and load any required packages for your analysis.

library(salesforcer)
sf_auth()

In the example below, we demonstrate how to create a record, but use the control arguments to prevent its creation from showing up in the Chatter feeds by setting the DisableFeedTrackingHeader.

new_contact <- c(FirstName = "Jenny", 
                 LastName = "Williams", 
                 Email = "jennyw@gmail.com")
record1 <- sf_create(new_contact,
                     object_name = "Contact",
                     DisableFeedTrackingHeader = list(disableFeedTracking = TRUE))
record1
#> # A tibble: 1 x 2
#>   success errors    
#>   <lgl>   <list>    
#> 1 FALSE   <list [1]>

You will notice that the argument DisableFeedTrackingHeader can be included right into the function without any documentation existing for it in the sf_create function. This is because the dots (...) allow you to pass over a dozen different control parameters and that documentation would be tedious to create and maintain over multiple functions in the package. However, you will notice in the documentation entry for the control argument there is a link to a function called sf_control which you can use to directly to pass into control or simply to review its documentation of all the possible control parameters and their defaults. This is where you can review the various control options in more detail before trying to set them.

You may have also noticed that the argument DisableFeedTrackingHeader was formatted as a list with an element inside called disableFeedTracking set to TRUE. This may seem redundant but there are two reasons for this. First, this is exactly how the Salesforce APIs documents these options, which are typically referred to as “headers” because they are passed as a named header of the HTTP request and then the header fields and values are provided for that header. Second, some headers have multiple fields and values so a list is the only way to provide multiple named fields and values under a single header entity.

An example using the DuplicateRuleHeader

The DuplicateRuleHeader that controls whether the duplicate rules are followed when inserting records from the API, has three fields:

  1. allowSave - For a duplicate rule, when the Alert option is enabled, bypass alerts and save duplicate records by setting this property to true. Prevent duplicate records from being saved by setting this property to false.

  2. includeRecordDetails - Get fields and values for records detected as duplicates by setting this property to true. Get only record IDs for records detected as duplicates by setting this property to false.

  3. runAsCurrentUser - Make sure that sharing rules for the current user are enforced when duplicate rules run by setting this property to true. Use the sharing rules specified in the class for the request by setting this property to false. If no sharing rules are specified, Apex code runs in system context and sharing rules for the current user are not enforced.

Specifying these arguments requires a list structure in R, which may seem redundant in some cases, but is necessary to follow in order to build the API request correctly.

# override the duplicate rules ...
record2 <- sf_create(new_contact,
                     object_name = "Contact",
                     DuplicateRuleHeader = list(allowSave = TRUE, 
                                                includeRecordDetails = FALSE, 
                                                runAsCurrentUser = TRUE))
record2
#> # A tibble: 1 x 2
#>   id                 success
#>   <chr>              <lgl>  
#> 1 0033s000017uOvnAAE TRUE

# ... or succumb to the duplicate rules
record3 <- sf_create(new_contact,
                     object_name = "Contact",
                     DuplicateRuleHeader = list(allowSave = FALSE, 
                                                includeRecordDetails = FALSE, 
                                                runAsCurrentUser = TRUE))
record3
#> # A tibble: 1 x 2
#>   success errors    
#>   <lgl>   <list>    
#> 1 FALSE   <list [1]>

Per the description above, note that setting allowSave=TRUE will not override rules where the “Action on Create” for a rule is set to “Block”. If the duplicate rule’s action is “Allow” with an alert, then setting allowSave=TRUE means the record will be created with no warning message. If allowSave=FALSE, then the record will be prevented from being created. For additional information on the DuplicateRuleHeader, please see the Salesforce documentation at: https://developer.salesforce.com/docs/atlas.en-us.api.meta/api/sforce_api_header_duplicateruleheader.htm

Finally, you may notice during your use that only certain control arguments are permitted based on the API. For example, the DuplicateRuleHeader is not implemented in the REST API like it is in the SOAP API. In the example below you should take note of two things:

  1. When using the REST API and setting the DuplicateRuleHeader, then you will notice a warning that the header was ignored. You will receive warnings when trying to set any control parameters for an API or operation that does not recognize that particular control.

  2. In this example, you cannot bypass the duplicate rule alert to create the record if using the REST API like you can with the SOAP API.

record4 <- sf_create(new_contact,
                     object_name = "Contact",
                     DuplicateRuleHeader = list(allowSave = FALSE, 
                                                includeRecordDetails = FALSE, 
                                                runAsCurrentUser = TRUE),
                     api_type = "REST")
#> Warning: Ignoring the following controls which are not used in the REST API:
#> DuplicateRuleHeader
record4
#> # A tibble: 1 x 2
#>   success errors    
#>   <lgl>   <list>    
#> 1 FALSE   <list [1]>

Creating the control argument with sf_control

If this type of control structure is new to you, take a look at the documentation for the glm and glm.control functions. The way these two functions behave is exactly how functions like sf_create and sf_control work with each other. As demonstrated above you can pass any number of arbitrary controls into the function and they are all gathered up into the control by control = list(...). However, you can specify the control directly like this:

sf_query("SELECT Id, Name FROM Account LIMIT 1000",
         object_name = "Account",
         control = sf_control(QueryOptions = list(batchSize = 200)))
#> # A tibble: 15 x 2
#>   Id                 Name                                
#>   <chr>              <chr>                               
#> 1 0013s00000zFgA6AAK KEEP Test Account With Child Records
#> 2 0013s00000zFdugAAC KEEP Test Account With Child Records
#> 3 0013s000014jF2vAAE Test Account For Performance Test   
#> 4 0016A0000035mJEQAY GenePoint                           
#> 5 0016A0000035mJCQAY United Oil & Gas, UK                
#> # … with 10 more rows

Backwards compatibility for all_or_none and other named arguments

You may already be taking advantage of the all_or_none or line_ending arguments which are control arguments that were explicity included in functions. These argument essentially hard coded values to pass the AllOrNoneHeader and LineEndingHeader control parameters. Starting with the 0.1.3 release it is no longer necessary and preferable not to have an argument like all_or_none listed explicity as an argument since it can be provided in the control argument. Note: the all_or_none argument and other explicit control arguments will still be available in {salesforcer} 0.1.3 but will provide a deprecated warning. They will be removed in the next CRAN release of the package so it will be important to update your code now if you are explicitly passing these arguments and see a deprecation warning.